J Neurol Neurosurg Psychiatry: first published as 10.1136/jnnp.49.10.1142 on 1 October 1986. Downloaded from

Journal of Neurology, Neurosurgery, and Psychiatry 1986;49:1142-1145

Carbamazepine-viloxazine interaction in patients with

F PISANI, A FAZIO, G OTERI, E PERUCCA, M RUSSO, R TRIO, B PISANI, R DI PERRI From the Institute ofNeurological and Neurosurgical Sciences, University ofMessina, Messina and Institute ofMedical Pharmacology, University ofPavia, Pavia, Italy

SUMMARY In six depressed epileptic patients stabilised on therapy, addition of the antidepressant agent viloxazine (300 mg/day for three weeks) induced a marked (average 55%) increase in steady-state plasma carbamazepine concentration. The concentration of the active metabolite carbamazepine-10,11 -epoxide also increased during viloxazine therapy, but to a lesser extent (16%). In three patients, these effects were associated with symptoms of carbamazepine intoxication, which regressed rapidly when plasma carbamazepine and carbamazepine-10,11- epoxide levels returned to baseline values after discontinuation of viloxazine. In a seventh patient, viloxazine had to be discontinued after only two weeks because of severe side effects associated with guest. Protected by copyright. a striking elevation of carbamazepine and carbamazepine 10,11 -epoxide levels (by 197% and 137% respectively). Although viloxazine appears to be one of the few antidepressants which can be used safely in patients with epilepsy these results indicate that the drug should be prescribed with great caution in subjects treated with carbamazepine. The mechanism of the interaction probably involves inhibition of the of both carbamazepine and its active epoxide metabolite.

Viloxazine is a "second generation" antidepressant Material and methods which shows properties in some ani- mal models of epilepsy.' A number of studies, Seven hospitalised epileptic patients with depressive symp- recently reviewed by Edwards,2 3 provide evidence toms who had been on a constant dosage of carbamazepine that clinically the drug is non-epileptogenic or at least for at least 2 months agreed to take part in the trial (table 1). less than conventional antidepressants Viloxazine was administered for 3 weeks on a thrice daily epileptogenic regimen (100 mg at 9.00, 13.00 and 18.00). The dosage of and therefore could be used safely in depressed epi- CBZ was maintained unchanged throughout the study. leptic patients. A rational use of viloxazine in these Plasma carbamazepine and carbamazepine-E levels were patients, however, requires knowledge of its possible determined at weekly intervals before, during and after vil- interactions with concurrently administered anti- oxazine treatment. Blood samples were collected in hepa-

convulsants.4 rinised tubes before the first daily carbamazepine dose (9.00) http://jnnp.bmj.com/ In a previous paper,5 we reported that the addition and 3 hours later. The plasma was separated immediately of viloxazine to carbamazepine therapy induces an and stored at - 20° C until analysis. Carbamazepine and increase in serum carbamazepine levels leading to the carbamazepine-E concentrations were determined by HPLC of carbamazepine intoxication in some according to Milhaly et al6 with minor modifications (results development given in the text are means of values at the two sampling patients. The present study was designed to character- times). Plasma viloxazine levels were determined by a GLC ise this interaction in greater detail. To this purpose, method7 developed in our laboratory. Plasma lev- the behaviour of the active metabolite els were determined by EMIT.

carbamazepine- 10,1 -epoxide (carbamazepine E) was Statistical analysis was carried out by means of the Stu- on September 27, 2021 by also investigated. dent's t test for paired data. Address for reprint requests: Dr Francesco Pisani, Clinica Neuro- Results logica, Policlinico, 98013 Contesse-Messina, Italy. Received 31 December 1985. Plasma drug levels Six patients completed the Accepted 21 January 1986 3-week viloxazine treatment. In these patients, car- 1142 J Neurol Neurosurg Psychiatry: first published as 10.1136/jnnp.49.10.1142 on 1 October 1986. Downloaded from

Carbamazepine-viloxazine interaction in patients with epilepsy 1143

Table 1 Details ofpatients included in the trial Patient Sex Age (yr) Weight (kg) Seizure type* Seizure CBZ dosage Co- ftequencyt (mg/day) t I M 39 79 PC 2 600 - 2 M 39 80 PC 8 800 (4 mg/day) 3 M 48 83 PCSG 6 1200 (15 mg/day) 4 F 40 40 PC 12 600 - 5 F 19 55 PC 4 600 - 6 F 25 60 PC 10 600 - 7 M 33 71 PCSG 2 1300 Phenytoin (300 mg/day) *PC: partial complex; PCSG: partial complex with secondary generalization. tNumber of seizures in the week before viloxazine. $CBZ was taken in three divided daily doses. Table 2 Plasma CBZ and CBZ-E concentrations (pg/ml) in 7patients before, during (+) and after combined treatment with viloxazine (300 mg/day). Values shown are means oftwo determinations (each in duplicate) at 9.00 (before thefirst CBZ dose) and three hours later. Patient 1 2 3 4 5 6 7* Study day CHZ CBZ-E CBZ CBZ-E CBZ CBZ-E CBZ CBZ-E CBZ CBZ-E CBZ CBZ-E CBZ CBZ-E 1 553 051 47 048 128 1-33 76 078 58 065 6-4 070 7-4 1-67 7 52 056 50 047 13-6 1*38 8-1 0-80 6-2 070 6-8 0-67 7-2 1 54 14+ 7.3 0-57 8-8 0-55 19 3 1 58 10.9 0-86 8-4 0-82 9-6 0-81 10-3 1-64 21+ 9-0 0-59 7-9 0 52 20-4 1-66 11-2 0-89 9.1 0-78 10-8 0-86 21 4 3 55 28+ 92 059 75 050 208 162 117 090 90 075 111 091 77 195 35 5-5 0-51 45 0-48 118 1 31 72 071 54 0-62 59 0-63 74 1-65

42 50 053 53 047 122 1-37 76 074 58 0-68 6-2 065 - - guest. Protected by copyright. *Viloxazine was discontinued after two weeks in this patient because of severe intoxication.

bamazepine levels increased by 55% from 7 5 + 3-2 increase was less marked than that observed for the ig/ml (mean + SD) before viloxazine to 11-6 + 4-8 parent drug, the carbamazepine-E/CBZ ratio ug/ml after 3 weeks of viloxazine (p < 0 001), max- declined from 0-010 + 0-007 to 0-0075 + 0-008, that imal values being reached usually in the second or is 25% on average (p < 0 001) (fig 1). third week of associated treatment (table 2). In the seventh patient, viloxazine caused a striking Carbamazepine-E levels also increased from 0-76 + increase in both carbamazepine and carbamazepine-E 0-32 pg/ml to 0-88 + 040 (p < 0001): since this levels (from 7-2 to 21[4 and from 1 54 to 3-65 pg/ml, that is 197% and 137% respectively) and the anti- °--o CBZ-E 0-* CBZ *-- CBZ-E depressant had to be discontinued after two weeks CBZ _ I because of severe CBZ intoxication (fig 2). In this patient, the plasma levels of concurrently adminis- IL- CA tered phenytoin did not show any appreciable change during viloxazine treatment (range of observed val- ues: 12 3 to 16-1 pg/ml). In all patients, car- u 8- bamazepine and carbamazepine-E levels returned to the basal (pre-viloxazine) values when the anti- 12 1 depressant was stopped. http://jnnp.bmj.com/ a Plasma viloxazine concentrations showed wide 1. interindividual and intraindividual variations: 379 + 4- 112 ng/ml in the first morning samples and 1726 + 0. C6 Vdo(3 ay) 574 ng/ml in the sample collected 3 hours later. w CBZ Clinical findings During viloxazine treatment N patients 3, 4 and 6 developed mild symptoms consis- .M tent with carbamazepine intoxication (dizziness, 0 1 2 3 4 5 6 drowsiness, fatigue). These symptoms disappeared on September 27, 2021 by Weeks rapidly after withdrawal of the antidepressant. Patient 7 marked confusion, Fig 1 Effect ofviloxazine treatment on plasma developed drowsiness, carbamazepine (CBZ), carbamazepine-10,11-epoxide ataxia and dizziness; the severity of these symptoms (CBZ-E) and CBZ-E/CBZ ratio (means + SD) in six required withdrawal of this patient from the trial on patients. the second week. J Neurol Neurosurg Psychiatry: first published as 10.1136/jnnp.49.10.1142 on 1 October 1986. Downloaded from

1144 Pisani, Fazio, Oteri, Perucca, Russo, Trio, Pisani, Perri °--O CBZ-E *-* CBZ *-u CBZ-E availability but in view of the magnitude of the inter- CBZ action and the concurrent changes in metabolite levels 24- an interference with carbamazepine elimination -9 appears more likely. Carbamazepine is cleared almost cJ) entirely by biotransformation, the major pathways 1- being in order of importance epoxidation to mN . carbamazepine-E, hydroxylation of the aromatic u 16 rings and N-glucuronidation of the carbamoyl side

E chain.'0 Carbamazepine-E, in turn, is almost entirely Un *30 converted by the enzyme epoxide hydrolase to a a trans-diol derivative which is excreted in urine. Our results provide evidence that viloxazine interferes 8 8 with carbamazepine metabolism in a complex way: 15 0 the increase in the levels of parent drug is likely to result from inhibition of one or more of the main CBZ mu Vilox. (300mg/day) metabolic pathways of carbamazepine itself, while the CO U O. simultaneous elevation of the levels of the metabolite O 1 2 3 4 5 6 almost certainly reflects a viloxazine-induced Weeks inhibition of epoxide hydrolase. In view of the possi- Fig 2 Effect of viloxazine treatment on plasma bility that the rate of formation of the epoxide was carbamazepine (CBZ), carbamazepine-10,11-epoxide also reduced by viloxazine, the moderate elevation of (CBZ-E) and CBZ-E/CBZ ratio in patient 7. plasma carbamazepine-E may well underestimate the actual degree of epoxide hydrolase inhibition. Since None of the patients showed an increase in fit fre- this enzyme plays an important rolc in the guest. Protected by copyright. quency during the trial. A reduction in number of detoxification of reactive metabolites," further stud- seizures was seen in patients 3 and 6 during viloxazine ies to characterise specifically the effect of viloxazine treatment. Patient 3 had six seizures in the week on epoxide hydrolase would be desirable. before viloxazine, four, three and four seizures From the clinical point of view, the interaction respectively during the three weeks on viloxazine and described in this paper has important therapeutic six and five seizures respectively during the two weeks implications. Unless plasma carbamazepine levels are after viloxazine. In patient 6 the corresponding num- carefully monitored and carbamazepine dosage is bers of seizures were 10 (before viloxazine), six, four adjusted accordingly, the rise in anticonvulsant drug and four (during viloxazine) and nine, nine (after vil- levels could easily result in symptoms of intoxication. oxazine). The possible contribution of the concurrent elevation of plasma CBZ-E to the side effects observed in our Discussion patients is unclear, although it should be noted that high levels of carbamazepine-E (usually above those Although depression is a frequently encountered dis- seen in this study) have been implicated in the devel- order among epileptic patients,8 its treatment may be opment of clinical toxicity in patients receiving a com- troublesome because of the facilitating action of bination of carbamazepine and valpromide,'2 13 a many antidepressant drugs on seizure activity.3 Vil- potent inhibitor of epoxide hydrolase." Apart from oxazine appears to be an atypical antidepressant in the toxicological implications, the rise in plasma car- this respect because it exerts anticonvulsant effects in bamazepine and carbamazepine-E levels, together http://jnnp.bmj.com/ a number of animal models, even though at high with a possible anticonvulsant effect of viloxazine doses a proconvulsant action may be seen.' Available itself, could explain the improvement in seizure con- evidence from clinical studies suggests that viloxazine trol which has been reported in some depressed epi- has a low potential for precipitating seizures in leptic patients (including two from our series) when depressed subjects2 3 and might even improve seizure viloxazine was added to pre-existing medication.3 9 control in some epileptic patients.39 In the light of In the light of the evidence presented above, these data, an evaluation of the possible interactions frequent monitoring of plasma anticonvulsant drug with concomitantly administered antiepileptic drugs levels and close clinical observation are required on September 27, 2021 by is clearly warranted. whenever viloxazine is prescribed in these patients. The present study demonstrates a clinically important interaction with carbamazepine. The- Mrs Lorefice Carmela is thanked for her technical oretically, all increase in plasma carbamazepine levels assistance and the Italian ICI Company for its gener- could result from enhancement of its oral bio- ous financial support. J Neurol Neurosurg Psychiatry: first published as 10.1136/jnnp.49.10.1142 on 1 October 1986. Downloaded from

Carbamazepine-viloxazine interaction in patients with epilepsy 1145 References D'Agostino A, Oteri G, Pisani F. A sensitive gas- chromatographic assay for the determination of serum Meldrum BS, Anlerzark GM, Greenwood DT. Anti- viloxazine concentration using a nitrogen-phosphorus convulsant and proconvulsant properties of viloxazine selective detector. Therapeutic Drug Monitoring hydrochloride: pharmacological and pharmacokinetic 19$4;6:484-8. studies in rodents and the epileptic baboon. Psycho- 8 Betts TA. Depression, anxiety and epilepsy. In: Reynolds pharmacology (Berlin) 1982;76:212-7. EH, Trimble MR (eds). Epilepsy and Psychiatry, 2Edwards JG, Glen-Bott M. Does viloxazine have epi- Edinburgh: Churchill-Livingstone, 1981:60-71. leptogenic properties? J Neurol Neurosurg Psychiatry 9Pisani F, Narbone MC, Oteri G, Primerano G, Amendola 1984;47:960-4. D'Agnostino A, Fazio A, Nasso C, Romano F, Di Perri 3 Edwards JG. Antidepressants and seizures: epi- R. Influenza della viloxazine sulla frequenza delle crisi demiological and clinical aspects. In: Trimble MR (ed). epilettiche. Boll Lega It Epil 1985;49:231-2. The Psychopharmacology ofEpilepsy, Chichester: John Perucca E, Richens A. Clinical pharmacokinetics of anti- Wiley 1985:119-39. epileptic drugs. In: Frey HH, Janz D (eds). Antiepileptic 4Perucca E, Manzo L, Crema A. Pharmacokinetic inter- Drugs, Handbook of Experimental Pharmacology, actions between antiepileptic and psychotropic drugs. Vol. 74, Berlin, Springer-Verlag 1985:661-723. In: Trimble MR (ed). The Psychopharmacology of Pacifici GM, Tomson T, Bertilsson L, Rane A. Epilepsy, Chichester: John Wiley 1985:95-105. Valpromide/carbamazepine and risk of teratogenicity. 'Pisani F, Narbone MC, Fazio A, Crisafulli P, Primerano Lancet 1985;ii:397-8. G, Amendola D'agostino A, Oteri G, Di Perri R. Effect "Meijer JWA, Binnie CD, Debets RMChr, Van Parys ofviloxazine on serum carbamazepine levels in epileptic JAP, De Beer-Pawlikowski NKB. Possible hazard of patients. Epilepsia 1984;25:482-5. valpromide-carbamazepine in epilepsy. Lancet 1984;i: 6Mihaly GW, Phillips JA, Louis WJ, Vajda FJ. Mea- 802. surement of carbamazepine and its epoxide metabolite 13Pisani F, Fazio A, Ruello C, Gitto C, Russo F, Perucca E. by high-performance liquid chromatography, and a Differential interactions of valproic acid and comparison of assay techniques for the analysis of car- valpromide with carbamazepine in man. Proceedings guest. Protected by copyright. bamazepine. Clin Chem 1977;23:2283-7. 16th Epilepsy International Symposium, Hamburg 'Fazio A, Crisafulli P, Primerano G, Amendola September 6-9, 1985 (in press). http://jnnp.bmj.com/ on September 27, 2021 by